Unable to run the code by following the instruction

[ycliu93]

Thanks for providing interesting works and publicly releasing the code implementation.

I tried to follow the instructions and run the code, but I encountered some obstacles and got stuck (in stage 1).

1. mmcv version: since the current default version is 1.7.0 (I then tried to downgrade to 1.5.0)
2. DeRy/blocksize/block_meta.py does not have MODEL2MODULES, and this leads to the import error in DeRy/blocksize/__init__.py.
3. DeRy/mmcls and mmls package are conflicted, and renaming the DeRy/mmcls to other names can address this issue
4. DeRy/mmcls/datasets/multi_label.py is missing, and I tried to remove all parts that imported multi_label but some errors are still shown.
5. What should I assign for $Config for stage 1 (Model Zoo Preparation)?

Has anyone (or authors) successfully launched the training based on the current version?

This is an interesting work, and I would appreciate it if the authors can address some errors in the code and provide the launchable code 🙂

Thank you.

[Adamdad]

Dear @ycliu93 , Thank you so much for your interest in our work. I will address each of your question below:

1. I just comment out the mmcv version requirement in the mmcls/__init__.py. My code should be OK for up-to-date mmcv version. I ran all my experiments in mmcv-full==1.3.18, but it should not be a problem now.
2. I remove the variable name MODEL2MODULES in the current files. I used to adopt the MODEL2MODULES for my old code version, but I rename it to MODEL_BLOCKS now.
3. I rename the folder name to mmcls_addon. Because I used to be running the code by installing the mmcls from source code and directly edit the source code, I did not notice the conflict when I clean my code.
4. Use the configs in configs/compute_sim. They are the configs to load the pertained weights from mmcls or timm and compute the feature embeddings for the network partition.

So sorry for the inconvenience. Since I would like to make my code online as soon as possible. If you have any other questions, please feel free to contact me again.

By the way, I happened to be using your code for unbiased-teacher last year and find it very helpful. Thanks for your great work. 😊😊😊😊😊